基于情感字典与连词结合的中文文本情感分类
Classification of chinese texts sentiment based on semantic and conjunction
作者：刘玉娇(四川大学计算机学院)；琚生根(四川大学计算机学院)；伍少梅(四川大学计算机学院)；苏翀(四川大学计算机学院)
Author：LIU Yu-Jiao(College of Computer, Sichuan University)；JU Sheng-Gen(College of Computer, Sichuan University)；WU Shao-Mei(College of Computer, Sichuan University)；SU Chong(College of Computer, Sichuan University)
收稿日期：2014-03-28 年卷（期）页码：2015,52(1):57-62
期刊名称：四川大学学报: 自然科学版
Journal Name：Journal of Sichuan University (Natural Science Edition)
关键字：中文文本情感; 基准词; SO-PMI算法; 情感字典; 共现
Key words：Chinese text sentiment; Paradigm word; SO-PMI algorithm; Semantic lexicon; Co occurrence
基金项目：四川省科技支撑计划项目（2012GZ0091）；四川大学青年基金项目（2012SCU11033)
中文摘要
本文首先利用基础情感词典以及基准词对所需研究领域的评论文本进行分析, 以此获得特定领域具有感情倾向的特征词语. 而后利用基准词以及获得的特征词语对评论进行分析, 对于有感情词的句子, 采用计算感情值来判别其感情倾向以及感情程度, 对于无感情词的句子, 采用连词的方法来进行感情的判别. 实验结果表明, 该方法能够对不同领域的评论得到较好的感情分类效果.
英文摘要
This paper proposed a method to get features with emotional tendencies in specific areas by means of the basic emotional lexicon and paradigm words ,which was used for analyzing comments. In this paper, emotional value will be computed to get the emotional tendencies and level if sentences have emotional words.,and the sentences without emotional words will be judged by conjunction. The experimental results show that the proposed method can get good classification effect in different areas.
【关闭】
